[BUG] intel-quickspi: touchscreen DMA write timeout (Goodix 27C6:012D) on Lenovo YOGA Pro 16 IAH10

[BUG] intel-quickspi: touchscreen DMA write timeout (Goodix 27C6:012D) on Lenovo YOGA Pro 16 IAH10

[Qiao Ruihao]

Hello,

The touchscreen on my Lenovo YOGA Pro 16 IAH10 does not work under
Linux. The intel-quickspi driver detects the device, but the
feature-report fetch (GET_REPORT) fails with a TxDMA write timeout,
which makes hid-multitouch fail initialization and fall back to a
generic keyboard, so the screen does not respond to any touch input.

The touchscreen works correctly under Windows 11 on the same hardware.

## Hardware

  - Laptop: Lenovo YOGA Pro 16 IAH10 (MT 83L0)
  - CPU: Intel Arrow Lake-H
  - Touchscreen: Goodix 27C6:012D, connected via Intel THC QuickSPI
  - THC controller: 00:10.0 Digitizer Pen [8086:7749] (rev 03),
    subsystem Lenovo [17aa:3815]

## Software

  - Kernel: 7.0.14-070014-generic
    (also reproduced on 6.19.14-061914-generic)
  - Distro: Ubuntu 22.04.5 LTS

## Problem

At boot the device is detected (reset + device descriptor read
succeed), but then hid-multitouch fails to fetch its feature reports
because every GET_REPORT command sent over TxDMA times out.

## Relevant kernel log

  intel_quickspi 0000:00:10.0: Can't find wake GPIO resource
  ACPI: \_SB.PC00.THC0._RST: Excess arguments - Caller passed 1, ACPI requires 0
  input: quickspi-hid 27C6:012D Touchscreen as /devices/pci0000:00/0000:00:10.0/0001:27C6:012D.0002/input/input7
  intel_quickspi 0000:00:10.0: Wait Get Report Response timeout, ret:0
  hid-multitouch 0001:27C6:012D.0002: failed to fetch feature 2
  intel_quickspi 0000:00:10.0: timeout for waiting TxDMA completion
  intel_quickspi 0000:00:10.0: DMA write failed, ret = -110
  intel_quickspi 0000:00:10.0: Write GET_REPORT command failed, ret = -110
  hid-multitouch 0001:27C6:012D.0002: failed to fetch feature 3
  hid-multitouch 0001:27C6:012D.0002: input,hidraw1: <UNKNOWN> HID v8.65 Keyboard [quickspi-hid 27C6:012D]

## Analysis

The reset (acpi_tic_reset via _RST) and device-descriptor read
succeed, so the SPI link and interrupt path partially work. The
failure is specific to quickspi_get_report() ->
write_cmd_to_txdma() -> thc_dma_write(): the TxDMA transfer never
completes (write_done is never set, 1s timeout -> -ETIMEDOUT).

I would be happy to provide additional information (full dmesg, ACPI
tables, register dumps) or test patches.

Thanks,
Qiao Ruihao

Re: [BUG] intel-quickspi: touchscreen DMA write timeout (Goodix 27C6:012D) on Lenovo YOGA Pro 16 IAH10

[Lovekesh Solanki]

On Sat, Aug 15, 2026 at 03:19:35PM +0000, Qiao Ruihao wrote:
 
> The touchscreen on my Lenovo YOGA Pro 16 IAH10 does not work under
> Linux. The intel-quickspi driver detects the device, but the
> feature-report fetch (GET_REPORT) fails with a TxDMA write timeout,
> which makes hid-multitouch fail initialization and fall back to a
> generic keyboard, so the screen does not respond to any touch input.

Could you attach full dmesg too?

RE: [BUG] intel-quickspi: touchscreen DMA write timeout (Goodix 27C6:012D) on Lenovo YOGA Pro 16 IAH10

[Xu, Even]

Generally speaking, a write timeout does not normally occur on the SPI bus. 
If it does happen, it usually indicates that there is a problem with the bus, or that the touch device has lost power. 
You can try unloading the intel_quickspi driver with rmmod, then reload it with insmod, and check the kernel log to see whether a different symptom appears.

Best Regards,
Even Xu

> -----Original Message-----
> From: Qiao Ruihao <qiaoruihao20061203@qq.com>
> Sent: Saturday, August 15, 2026 11:20 PM
> To: linux-input@vger.kernel.org
> Cc: Xu, Even <even.xu@intel.com>; Sun, Xinpeng <xinpeng.sun@intel.com>
> Subject: [BUG] intel-quickspi: touchscreen DMA write timeout (Goodix
> 27C6:012D) on Lenovo YOGA Pro 16 IAH10
> 
> Hello,
> 
> The touchscreen on my Lenovo YOGA Pro 16 IAH10 does not work under Linux.
> The intel-quickspi driver detects the device, but the feature-report fetch
> (GET_REPORT) fails with a TxDMA write timeout, which makes hid-multitouch fail
> initialization and fall back to a generic keyboard, so the screen does not respond
> to any touch input.
> 
> The touchscreen works correctly under Windows 11 on the same hardware.
> 
> ## Hardware
> 
>   - Laptop: Lenovo YOGA Pro 16 IAH10 (MT 83L0)
>   - CPU: Intel Arrow Lake-H
>   - Touchscreen: Goodix 27C6:012D, connected via Intel THC QuickSPI
>   - THC controller: 00:10.0 Digitizer Pen [8086:7749] (rev 03),
>     subsystem Lenovo [17aa:3815]
> 
> ## Software
> 
>   - Kernel: 7.0.14-070014-generic
>     (also reproduced on 6.19.14-061914-generic)
>   - Distro: Ubuntu 22.04.5 LTS
> 
> ## Problem
> 
> At boot the device is detected (reset + device descriptor read succeed), but then
> hid-multitouch fails to fetch its feature reports because every GET_REPORT
> command sent over TxDMA times out.
> 
> ## Relevant kernel log
> 
>   intel_quickspi 0000:00:10.0: Can't find wake GPIO resource
>   ACPI: \_SB.PC00.THC0._RST: Excess arguments - Caller passed 1, ACPI requires 0
>   input: quickspi-hid 27C6:012D Touchscreen as
> /devices/pci0000:00/0000:00:10.0/0001:27C6:012D.0002/input/input7
>   intel_quickspi 0000:00:10.0: Wait Get Report Response timeout, ret:0
>   hid-multitouch 0001:27C6:012D.0002: failed to fetch feature 2
>   intel_quickspi 0000:00:10.0: timeout for waiting TxDMA completion
>   intel_quickspi 0000:00:10.0: DMA write failed, ret = -110
>   intel_quickspi 0000:00:10.0: Write GET_REPORT command failed, ret = -110
>   hid-multitouch 0001:27C6:012D.0002: failed to fetch feature 3
>   hid-multitouch 0001:27C6:012D.0002: input,hidraw1: <UNKNOWN> HID v8.65
> Keyboard [quickspi-hid 27C6:012D]
> 
> ## Analysis
> 
> The reset (acpi_tic_reset via _RST) and device-descriptor read succeed, so the
> SPI link and interrupt path partially work. The failure is specific to
> quickspi_get_report() ->
> write_cmd_to_txdma() -> thc_dma_write(): the TxDMA transfer never completes
> (write_done is never set, 1s timeout -> -ETIMEDOUT).
> 
> I would be happy to provide additional information (full dmesg, ACPI tables,
> register dumps) or test patches.
> 
> Thanks,
> Qiao Ruihao